DE102022004349A1 - Process for driver assistance and vehicle - Google Patents

Abstract

The invention relates to a method for driver assistance in a vehicle (1) with a control unit (2), a further control unit (3) for a driver assistance system, sensors (5) for measuring precipitation, lighting conditions and temperature, at least one camera (6) and a data transmission (7) to a cloud (8), with the sensors (5), the LiDAR sensor (4) and the camera (6) detecting an area surrounding the vehicle (1) while driving, with the detected sensor information being stored in are aggregated and recorded in the control device (2), whereby upon detection of a game crossing or a danger zone with an increased risk of deer crossing using the sensor information and marking a geoposition from all sensor information in the control unit (2) and/or in the cloud (8) using a machine method Learning, a predictive model of the risk of deer crossing is generated, with the presence of patterns in the sensor information, which the predictive model has an increased G In the event of a deer crossing (10), a visual, acoustic and/or haptic warning, for example, is issued to a driver of the vehicle (1), with detected probable danger spots being sent to the cloud (8) for aggregation of the information and transmission of the aggregated information to a vehicle fleet be sent.

Description

The invention relates to a driver assistance method according to the preamble of claim 1 and a vehicle according to the preamble of claim 7.

• - anhand der erfassten Signale ein Tier neben oder auf einer Fahrbahn des jeweiligen Fahrzeuges detektiert wird und eine Information über eine erkannte Tierart, eine erkannte Trajektorie des Tieres, eine Bewegungsgeschwindigkeit des Tieres sowie ein Ort und eine Uhrzeit der Detektion des Tieres gespeichert und an eine mit dem jeweiligen Fahrzeug datentechnisch gekoppelte zentrale Rechnereinheit übermittelt wird,
• - diese der zentralen Rechnereinheit jeweils übermittelten ortsabhängigen Informationen aggregiert werden und wenn eine Anzahl übermittelter Informationen für neben oder auf dieser Fahrbahn detektierte Tiere einen vorgegebenen Schwellwert überschreitet, ermittelt wird, dass es sich um eine Tierfährte handelt, und
• - bei Befahren der Fahrbahn durch ein Fahrzeug der Fahrzeugflotte ein Warnhinweis in Bezug auf ein Vorhandensein einer Tierfährte in dem die Fahrbahn befahrenden Fahrzeug ausgegeben wird und/oder eine Fahrgeschwindigkeit des Fahrzeuges automatisch verringert wird.

DE 10 2022 002 339 A1 describes a method for configuring and operating a deer crossing warning system of a vehicle in a vehicle fleet with an environment sensor system which continuously detects signals, at least when the vehicle is driving, by means of which an environment of the vehicle and objects located in it are detected. It is provided that

• - Based on the detected signals, an animal is detected next to or on a roadway of the respective vehicle and information about a recognized animal species, a recognized trajectory of the animal, a movement speed of the animal and a location and time of detection of the animal is stored and sent to a the respective vehicle is transmitted to a data-technically coupled central computer unit,
• - this location-dependent information transmitted to the central computer unit is aggregated and if a number of transmitted information for animals detected next to or on this roadway exceeds a predetermined threshold value, it is determined that it is an animal track, and
• - When driving on the roadway by a vehicle of the vehicle fleet, a warning regarding the presence of an animal track in the vehicle driving on the roadway is issued and/or a driving speed of the vehicle is automatically reduced.

The invention is based on the object of specifying a novel method for driver assistance for warning of deer crossings and a novel vehicle with driver assistance for warning of deer crossings.

The object is achieved according to the invention by a method for driver assistance with the features of claim 1 and by a vehicle with the features of claim 7.

Advantageous configurations of the invention are the subject matter of the dependent claims.

A driver assistance method according to the invention is carried out in a vehicle with a control unit, a further control unit for a driver assistance system, optionally a LiDAR sensor, further sensors for measuring precipitation, lighting conditions and temperature, at least one camera and data transmission to a cloud, with during During the journey, the sensors, the LiDAR sensor and the camera are used to capture the surroundings of the vehicle, with a deer crossing or a hazardous area with an increased risk of deer crossings being identified using the sensor information using a computer-assisted vision method. According to the invention, it is provided that detected sensor information is aggregated and recorded in the control unit, with the detection of a deer crossing or a hazardous area with an increased risk of deer crossing using the sensor information and marking a geoposition from all sensor information in the control unit and / or in the cloud using a machine learning method a predictive model of deer crossing awareness is generated, with the presence of patterns in the sensor information to which the predictive model assigns an increased risk of deer crossing, a warning, for example an optical, acoustic, haptic, and / or olfactory warning and / or a warning by means a brain-computer interface, is issued to a driver of the vehicle, with detected probable danger spots being sent to the cloud for aggregation of the information and transmission of the aggregated information to a vehicle fleet. The LiDAR sensor is an optional example of sensors that can be used to characterize the environment. For example, ultrasonic sensors or radar sensors are also suitable. The method according to the invention can also only be carried out with the aid of the at least one camera, in particular an external camera.

The solution according to the invention makes it possible to generate animal crossing warnings by predicting probable animal crossings in a vehicle and to recognize danger spots that were previously unknown. In this way, the vehicle is prepared for an occurring dangerous situation that has not been documented by animal crossings that have already occurred, been observed or evaluated.

The invention increases safety and driving comfort in a vehicle, since the vehicle can preventively react to future probable animal crossings and thus increase safety and driving comfort with digital means. Furthermore, the information about game crossing locations can be made available, for example, for scientific purposes or other purposes serving the public interest.

Exemplary embodiments of the invention are explained in more detail below with reference to drawings.

• 1 eine schematische Ansicht eines Fahrzeugs,
• 2 eine schematische Ansicht eines auf einer Fahrbahn mit einem Wildwechsel fahrenden Fahrzeugs,
• 3 eine weitere schematische Ansicht eines auf einer Fahrbahn mit einem Wildwechsel fahrenden Fahrzeugs,
• 4 eine weitere schematische Ansicht eines auf einer Fahrbahn mit einem Wildwechsel fahrenden Fahrzeugs, und
• 5 eine weitere schematische Ansicht eines auf einer Fahrbahn mit einem Wildwechsel fahrenden Fahrzeugs.

show:

• 1 a schematic view of a vehicle,
• 2 a schematic view of a vehicle driving on a roadway with a deer crossing,
• 3 a further schematic view of a vehicle driving on a roadway with a deer crossing,
• 4 a further schematic view of a vehicle driving on a roadway with a deer crossing, and
• 5 a further schematic view of a vehicle driving on a roadway with a deer crossing.

Corresponding parts are provided with the same reference symbols in all figures.

1 is a schematic view of a vehicle 1 with a control unit 2, a further control unit 3, for example for a driver assistance system (ADAS), a plurality of sensors 4, 5, 6, including in particular a LiDAR sensor 4, one or more further sensors 5, for example rain gauge, light meter, temperature sensor, etc., one or more cameras 6, a data transmission 7 and a cloud 8.

The surroundings of the vehicle 1 are detected with the sensors 4 to 6 .

In the vehicle 1, sensor information from the sensors 4, 5, 6 (e.g. the camera 6, the LiDAR sensor 4 and the temperature sensor) and information from the other control units 3 (e.g. ADAS, engine control unit, navigation unit) are aggregated and recorded in the control unit 2 .

If a deer crossing is recognized based on the sensor information or if a place with an increased danger of deer crossings is driven through, then a predictive model of the danger of deer crossings is generated in the control unit 2 from all sensor information, in particular recognized features and characteristic values.

For this purpose, the data transmission 7 to the cloud/computer center 8 can also be used in order to generate a predictive model of the risk of wildlife crossing from the identified features and characteristic values in the cloud/computer center 8 .

2 is a schematic view of a vehicle 1 driving on a roadway 9 with a deer crossing 10. 3 is a further schematic view of a vehicle 1 driving on a roadway 9 with a deer crossing 10.

Sensor information from sensors 4, 5, 6 is recorded in vehicle 1. If a deer crossing 10 is detected in the vehicle 1 and/or a location with an increased probability of a deer crossing 10, which was generated from aggregated fleet data, is driven over, this sensor information is aggregated in the control unit 2 and recorded. There or in the cloud 8, a predictive model of the risk of deer crossing is generated from this data.

4 is a schematic view of a vehicle 1 driving on a lane 9. 5 is a further schematic view of a vehicle 1 driving on a roadway 9.

In the vehicle 1, sensor information from the sensors 4, 5, 6 and information from the other control units 3 (for example ADAS, engine control unit) are continuously evaluated in the control unit 2 using the predictive model. If patterns are found in the sensor information, to which the predictive model assigns a high or at least increased risk of deer crossing 10 (i.e. with a probability of deer crossing 10 above a certain threshold value), then vehicle 1 issues a warning to the driver. In 4 For example, rock formations 11 to the left and right of the roadway 9 and a water point 12 are recognized. The predictive model has learned that such a spot is increasingly used as a transition from wild, i.e. wild crossing 10. In 5 a mountain formation 13 with a water point 12 is recognized on the left side of the roadway 9 and a free grassy landscape 14 on the right side of the roadway 9, which the game can use as a source of food.

DE 10 2022 002 339 A1 describes a method for configuring and operating a wildlife crossing warning system of a vehicle in a vehicle fleet with an environment sensor system, which continuously detects signals, at least while the vehicle is driving, by means of which an environment of the vehicle in question and objects located in this environment are detected. DE 10 2022 002 339 A1 is hereby fully incorporated by reference into the present application.

DE 10 2022 002 339 A1 describes, for example, the recognition of signs of deer crossing (e.g. traces of movement in the sand, snow and/or mud or animal hair on the roadside) or animals on the roadside using methods of computer-assisted vision. For this purpose, for example, sensor data is recorded in the vehicle 1 and objects and their positions are evaluated using suitable object recognition methods.

Furthermore, warnings of wildlife crossings 10 can be transmitted to the vehicle 1 by means of car2car communication, infrastructure2car communication, information in a navigation system about danger spots and/or by automatic evaluation of roadway signs.

If a warning of a deer crossing 10 is now generated in the vehicle 1 using the above-mentioned methods, a current geoposition is marked and the sensor information of the vehicle 1, for example a camera image, a time, a temperature, precipitation conditions, etc. is recorded.

The sensor information is evaluated in the vehicle 1 or transmitted to the cloud 8 in order to be aggregated there. The predictive model for predicting probable animal crossings 10 is generated in the vehicle 1 or in the cloud 8 using the current sensor information. For this purpose, the sensor information is evaluated using machine learning methods to identify patterns in the sensor information which can predict a deer crossing 10 with a high degree of probability.

In one embodiment, the predictive model can predict the time-dependent probability of a wildlife crossing 10 at a location. For this purpose, seasonal information, for example sunrise times, sunset times, rutting times, as well as migration directions (for example north or south) during a certain season, for example spring or autumn, can be combined with data from a vehicle's weather sensor system in order to further improve the forecast if deer crossing periods shift , for example due to certain outside temperatures, amount of snow, etc.

For example, it can be recognized as a pattern that depressions lying in the shade in the early hours of the morning in the months of March, April and October are associated with a high or at least increased probability of a deer crossing 10 .

In a further embodiment, parts of the predictive model or the improved predictive model can be generated in the vehicle 1 and only the evaluated improvements can be transmitted to the cloud 8 in order to increase the protection of driver data (e.g. via federated learning).

The selection of training data and/or a training method also makes it possible for new behavior to be learned quickly and old behavior to be slowly unlearned.

During the journey, the predictive model is used continuously to predict probable animal crossings 10 based on current sensor information in the vehicle 1 . For this purpose, the sensor information is compared with the patterns recognized as important in the sensor information.

If the predictive model predicts a deer crossing 10 with a high degree of probability above a predefined threshold value, an optical, acoustic or haptic warning is output to the driver, for example, and the vehicle is optionally prepared for an occurring dangerous situation (e.g. by reducing the speed, preparing control units for driver assistance systems, activation of a dashcam).

Identified probable danger spots can be sent to the cloud 8 for aggregation of the information and transmission of the aggregated information to a vehicle fleet.

Reference List

1

vehicle

2

control unit

3

another control unit

4

Sensor, LiDAR sensor

5

sensor, another sensor

6

sensor, camera

7

data transmission

8th

cloud

9

roadway

10

animal crossing

11

stone formation

12

water point

13

mountain formation

14

grassland

QUOTES INCLUDED IN DESCRIPTION

This list of the documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• DE 102022002339 A1 [0002, 0021, 0022]

Claims (7)

Method for driver assistance in a vehicle (1) with a control unit (2), a further control unit (3) for a driver assistance system, sensors (5) for measuring precipitation, lighting conditions and temperature, at least one camera (6) and data transmission (7 ) to a cloud (8), with the sensors (5) and the camera (6) recording the surroundings of the vehicle (1) while driving, with a deer crossing or a hazardous area with an increased risk of deer crossing using the sensor information using a method of computer-assisted vision is recognized, characterized in that detected sensor information is aggregated and recorded in the control unit (2), wherein upon detection of a deer crossing or a hazardous area with an increased risk of deer crossing using the sensor information and marking a geo-position from all the sensor information in the control unit (2) and/ or in the cloud (8) using a machine learning method the model of the danger of deer crossing is generated, with the presence of patterns in the sensor information to which the predictive model assigns an increased risk of deer crossing (10), for example a visual, acoustic and/or haptic warning being issued to a driver of the vehicle (1). , wherein detected probable danger spots are sent to the cloud (8) for aggregation of the information and transmission of the aggregated information to a vehicle fleet. procedure after claim 1 , characterized in that the vehicle (1) is prepared for an occurring hazardous situation when a probable hazardous area is detected. procedure after claim 2 , characterized in that when preparing for the dangerous situation, the speed is reduced and/or at least one control unit for driver assistance systems and/or a dashcam is activated. Method according to one of the preceding claims, characterized in that warnings of deer crossings (10) are transmitted to the vehicle (1) by means of car2car communication, infrastructure2car communication, information in a navigation system about danger spots and/or by automatic evaluation of roadway signs . Method according to one of the preceding claims, characterized in that the predictive model predicts the time-dependent probability of a deer crossing (10) at a location by seasonal information, in particular sunrise times, sunset times, rutting times, and migration directions of game during a specific season with data from a Weather sensors of the vehicle (1) are combined. Method according to one of the preceding claims, characterized in that parts of the predictive model are generated in the vehicle (1) and only evaluated improvements are transmitted to the cloud (8). Vehicle (1) with a control unit (2), a further control unit (3) for a driver assistance system, sensors (5) for measuring precipitation, lighting conditions and temperature, at least one camera (6) and data transmission (7) to a cloud ( 8), characterized in that the vehicle (1) is configured to carry out the method according to one of the preceding claims.

Images